Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method for detecting weak and small target by using deep learning method in complex traffic environment

A traffic environment and weak target technology, applied in the field of traffic target recognition, can solve problems such as difficult to meet actual needs, limit detection accuracy, and difficult to detect vehicles, etc., to achieve good practical generalization performance, stable recognition accuracy, and excellent The effect on actual generalization performance

Pending Publication Date: 2022-06-03
刘建芳
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] A vision-based vehicle detection system emerged as the times require. The visual vehicle detection system can detect vehicle types, traffic volume, vehicle speed and even predict traffic accidents. However, as a background difference method, the current targ

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Method for detecting weak and small target by using deep learning method in complex traffic environment
  • Method for detecting weak and small target by using deep learning method in complex traffic environment
  • Method for detecting weak and small target by using deep learning method in complex traffic environment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0118] Embodiment 1: S5. Target detection experiment:

[0119] S501. Experimental setup

[0120] (1) Use Pytorch and the mmdetection open source framework provided by the Chinese University of Hong Kong to conduct experiments under the Ubuntu16.04 system; the data set used in the experiment is the kitti public data set, which is one of the most commonly used data sets in the field of autonomous driving. One of the common visual evaluation algorithm datasets; this dataset contains pictures in multiple common scenes such as urban roads, residential areas, campuses, etc., including Car, Van, Truck, Pedestrian, Person(sitting), Cyclist , Tram and Mis have eight categories, and there are up to 15 vehicles in each picture; it is mainly for detection of vehicles in motor vehicle lanes and other weak objects (pedestrians, pets, etc.) in sidewalks, so Car, Van, Truck and Tram are merged into a class of vehicles, and other classes are merged into a class of weak and small targets to fo...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a method for detecting a weak and small target by using a deep learning method in a complex traffic environment, the method detects the weak and small target by using an improved Faster R-CNN convolutional neural network, can obtain more effective feature expression and rich semantic information, introduces a feature pyramid network in an RPN network to generate a target candidate frame, and improves the target detection accuracy. According to the method, a convolutional neural network structure is optimized, a candidate box with more effective information can be obtained, the expression ability of a feature region containing important information is enhanced, efficient use of a region containing a target in an image is realized, and experimental verification shows that the method has the characteristics of high accuracy and good actual generalization performance.

Description

technical field [0001] The present invention is in the technical field of traffic target recognition, in particular to a method for detecting weak and small targets by using a deep learning method in a complex traffic environment. Background technique [0002] Image is an important source of information for human beings, vision is one of the main ways to receive external information in daily life, and target detection is to use multi-directional knowledge such as image processing, pattern recognition, machine learning, etc. Objects of interest, combining target detection and recognition, is more difficult than image classification; in people's daily work and life, the intelligent transportation system collects road vehicle driving information through cameras, and then processes the information by the central computer. Realize the tracking and identification of vehicles, identify traffic illegal vehicles, assist in handling various traffic violations, reduce the work pressure...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06V20/52G06V20/58G06K9/62G06V10/25G06V10/764G06V10/774G06V10/82G06N3/04G06N3/08G08G1/01G08G1/017
CPCG06N3/08G08G1/0175G08G1/0116G06N3/047G06N3/045G06F18/2415G06F18/214
Inventor 刘建芳
Owner 刘建芳
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com